Abstract
Some recent work on conditional formulation of multivariate Gaussian Markov random fields is presented. The focus is on model constructions by compatible conditionals and coregionalization. Special attention is given to multivariate generalizations of univariate models. Beginning with univariate model constructions, a survey of key approaches to formulating multivariate extensions is presented. Two challenges in the formulation and implementation of multivariate models are highlighted: (1) entanglement of spatial and non-spatial components, and (2) enforcement for positivity condition. Managing the two challenges by decomposition, separation, and constrained parameterization is discussed. Also highlighted is the challenge of flexible modeling of (conditional) cross-spatial dependencies and, in particular, asymmetric cross-spatial dependencies. Interpretation of asymmetric cross-spatial dependencies is also discussed. A coregionalization framework which connects and unifies the various lines of model development is presented. The framework enables a systematic development of a broad range of models via linear and spatially varying coregionalization, respectively, with extensions to locally adaptive models. Formulation of multivariate models over variable-specific lattices is discussed. Selected models are illustrated with examples of Bayesian multivariate and spatiotemporal disease mapping. Potential applications of coregionalization models in imaging analysis, covariance modeling, dimension reduction, and latent variable analysis are briefly mentioned.
Similar content being viewed by others
References
Barnard J, McCulloch R, Meng X (2000) Modeling covariance matrices in terms of standard deviations and correlations, with application to shrinkage. Stat Sin 10:1281–1311
Banerjee S, Carlin PB, Gelfand AE (2014) Hierarchical modeling and analysis for spatial data, 2nd edn. Chapman & Hall, New York
Berman A, Plemmons RJ (1994) Nonnegative matrices in the mathematical sciences. SIAM, Philodelphia
Besag J (1974) Spatial interaction and the statistical analysis of lattice systems (with discussion). J R Stat Soc Ser B 36:192–236
Besag J, York J, Mollié A (1991) Bayesian image restoration, with two applications in spatial statistics. Ann Inst Stat Math 43:1–21
Besag J, Kooperberg C (1995) On conditional and intrinsic autoregressions. Biometrika 82(4):733–746
Botella-Rocamora P, Martinez-Beneito MA, Banerjee S (2015) A unifying modelling framework for highly multivariate disease mapping. Stat Med 34(9):1548–1559
Brezger A, Fahrmeir L, Hennerfeind A (2007) Adaptive Gaussian Markov random fields with applications in human brain mapping. Appl Stat 56(3):327–345
Brewer MJ, Nolan AJ (2007) Variable smoothing in Bayesian intrinsic autoregressions. Environmentrics 18:841–857
Brown PJ, Le ND, Zidek JV (1994) Multivariate spatial interpolation and exposure to air pollutants. Can J Stat 22(4):489–509
Carlin BP, Banerjee S (2003) Hierarchical multivariate CAR models for spatio-temporally correlated survival data (with discussion). In: Bernardo JM, Bayarri MJ, Berger JO, Dawid AP, Heckerman D, Smith AEM, West M (eds) Bayesian statistics 7. Oxford University Press, Oxford, pp 45–63
Congdon P (2008a) A spatial structural equation model for health outcomes. J Stat Plan Inference 138(7):2090–2105
Congdon P (2008b) A spatially adaptive conditional autoregressive prior for area health data. Stat Methodol 5(6):552–563
Conti S, O’Hagan A (2010) Bayesian emulation of complex multi-output and dynamic computer models. J Stat Plan Inference 140:640–651
Cressie N (1993) Statistics for spatial data, 2nd edn. Wiley, New York
Cressie N, Wikle CK (2011) Statistics for spatio-temporal data. Wiley, New York
Daniels MJ, Kass RE (1999) Nonconjugate Bayesian estimation of covariance matrices and its use in hierarchical models. J Am Stat Assoc 94(448):1254–263
Daniels MJ, Zhou Z, Zou H (2006) Conditionally specified spacetime models for multivariate processes. J Comput Graph Stat 15:157–177
Dawid AP (1981) Some matrix-variate distribution theory: notational considerations and a Bayesian application. Biometrika 68(1):265–274
Feingold DG, Varga RS (1962) Block diagonally dominant matrices and generalisations of the Gerschgorin circle theorem. Pac J Math 12:1241–1250
Furrer R, Sain SR (2010) spam: a sparse matrix R package with emphasis on MCMC methods for Gaussian Markov random fields. J Stat Softw 36(10):1–25
Genton MG (2007) Separable approximations of space–time covariance matrices. Environmentrics 18:681–695
Genton MG, Kleiber W (2015) Cross-covariance functions for multivariate geostatistics. Stat Sci 30(2):147–163
Gelfand AE, Ghosh SK (1998) Model choice: a minimum posterior predictive loss approach. Biometrika 85:111
Gelfand AE, Vounatsou P (2003) Proper multivariate conditional autoregressive models for spatial data analysis. Biostatistics 4:11–25
Gelfand AE, Schmidt AM, Banerjee S, Sirmans CF (2004) Nonstationary multivariate process modelling through spatially varying coregionalization (with discussion). Test 13:263–312
Gelman A, Meng XL, Stern H (1996) Posterior predictive assessment of model goodness-of-fit via realized discrepancies. Stat Sin 6:733807
Gneiting T, Raftery AE (2007) Strictly proper scoring rules, prediction, and estimation. J Am Stat Assoc 102(477):359–378
Goicoa T, Adin A, Ugarte MD, Hodges JS (2018) In spatio-temporal disease mapping models, identifiability constraints affect PQL and INLA results. Stoch Environ Res Risk Assess 32(2):749–770
Greco FP, Trivisano C (2009) A multivariate CAR model for improving the estimation of relative risks. Stat Med 28:1707–1724
Harville DA (2007) Matrix algebra from a statistician’s perspective. Springer, New York
Held L, Natrio I, Penton SE, Rue H, Becker N (2005) Towards joint disease mapping. Stat Methods Med Res 14:61–82
Hodges JS, Reich BJ (2010) Adding spatially-correlated errors can mess up the fixed effect you love. Am Stat 64(4):325–334
Hoff PD (2011) Separable covariance arrays via the Tucker product, with applications to multivariate relational data. Bayesian Anal 6(2):179–196
Jin X, Carlin BP, Banerjee S (2005) Generalized hierarchical multivariate CAR models for areal data. Biometrics 61:950–961
Jin X, Carlin BP, Banerjee S (2007) Order-free co-regionalized areal data models with application to multiple-disease mapping. J R Stat Soc Ser B 69(5):817–838
Kashyap RL, Chellappa R (1983) Estimation and choice of neighbors in spatial-interaction models of images. IEEE Trans Inf Theory IT–29(1):60–72
Kim H, Sun D, Tsutakawa RK (2001) A bivariate Bayes method for improving the estimates of mortality rates with a twofold conditional autoregressive model. J Am Stat Assoc 96(456):1506–1521
Knorr-Held L, Best NG (2000) Bayesian modelling of inseparable space–time variation in disease risk. Stat Med 19(17–18):2555–2567
Knorr-Held L, Best NG (2001) A shared component model for joint and selective clustering of two diseases. J R Stat Soc Ser A 164:73–85
Lawson A (2013) Bayesian disease mapping: hierarchical modeling in spatial epidemiology. Chapman & Hall, New York
Lee D (2011) A comparison of conditional autoregressive models used in Bayesian disease mapping. Spat Spatio-temporal Epidemiol 2(2):79–89
Lee D, Rushworth A, Sahu S (2014) A Bayesian localized conditional autoregressive model for estimating the health effects of air pollution. Biometrics 70:419–429
Leroux BG, Lei X, Breslow N (1999) Estimation of disease rates in small areas: a new mixed model for spatial dependence. In: Halloran ME, Berry D (eds) Statistical models in epidemiology, the environment and clinical trials. Springer, New York, pp 135–178
Lichstein JW, Simon TR, Shriner SA, Franzreb KE (2002) Spatial autocorrelation and autoregressive models in ecology. Ecol Monogr 72(3):445–463
Lindgren F, Rue H, Lindstrom J (2011) An explicit link between Gaussian fields and Gaussian Markov random field: the stochastic partial differential equation approach. J R Stat Soc B 73:423–498
Liu X, Wall MM, Hodges JS (2005) Generalized spatial structural equation models. Biostatistics 6(4):539–557
MacNab YC, Kemetic A, Gustafson P, Sheps S (2006) An innovative application of Bayesian disease mapping methods to patient safety research: a Canadian adverse medical event study. Stat Med 25:3960–3980
MacNab YC (2007) Spline smoothing in Bayesian disease mapping. Environmetrics 18:727–744
MacNab YC, Gustafson P (2007) Regression B-spline smoothing in Bayesian disease mapping: with an application to patient safety surveillance. Stat Med 26(24):4455–4474
MacNab YC (2009) Bayesian multivariate disease mapping and ecological regression with errors in covariates: Bayesian estimation of DALYs and ‘preventable’ DALYs. Stat Med 28(9):1369–1385
MacNab YC (2010) On Bayesian shared component disease mapping and ecological regression with errors in covariates. Stat Med 29:1239–1249
MacNab YC (2011) On Gaussian Markov random fields and Bayesian disease mapping. Stat Methods Med Res 20:49–68
MacNab YC (2014) On identification in Bayesian disease mapping and ecological–spatial regression. Stat Methods Med Res 23(2):134–55
MacNab YC (2016a) Linear models of coregionalization for multivariate lattice data: a general framework for coregionalized multivariate CAR models. Stat Med 35:3827–3850
MacNab YC (2016b) Linear models of coregionalization for multivariate lattice data: order-dependent and order-free MCARs. Stat Methods Med Res 25(4):1118–1144
Mardia KV (1988) Multi-dimensional multivariate Gaussian Markov random fields with application to image processing. J Multivar Anal 24:265–284
Mardia KV, Goodall CR (1993) Spatial–temporal analysis of multivariate environmental monitoring data. Multivar Environ Stat 6(76):347–385
Martinez-Beneito MA (2013) A general modelling framework for multivariate disease mapping. Biometrika 100(3):539–553
Martinez-Beneito MA, Botella-Rocamora P, Banerjee S (2017) Towards a multi-dimensional approach to Bayesian disease mapping. Bayesian Anal 12(1):239–259
Oliveira VD (2012) Bayesian analysis of conditional autoregressive models. Ann Inst Stat Math 64:107133
Pan JX, MacKenzie G (2003) On the modelling mean–covariance structures in longitudinal studies. Biometrika 90:239–244
Pan JX, MacKenzie G (2007) Modelling conditional covariance in the linear mixed model. Stat Model 7(1):49–71
Pourahmadi M (1999) Joint mean–covariance models with applications to longitudinal data: unconstrained parameterisation. Biometrika 86(3):677–690
Pourahmadi M (2000) Maximum likelihood estimation of generalised linear models for multivariate normal covariance matrix. Biometrika 87(2):425–435
Pourahmadi M (2007) Cholesky decompositions and estimation of a covariance matrix: orthogonality of variance-correlation parameters. Biometrika 94(4):1006–1013
Pourahmadi M (2011) Covariance estimation: the GLM and regularization perspectives. Stat Sci 26(3):369–387
Pourahmadi M (2013) High-dimensional covariance estimation. Wiley, New York
Pourmohamad T, Lee HKH (2016) Multivariate stochastic process models for correlated responses of mixed type. Bayesian Anal 11(3):797–820
R Development Core Team (2007) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna
Reich B, Hodges JS (2008) Modeling longitudinal spatial periodontal data: a spatially adaptive model with tools for specifying priors and checking fit. Biometrics 64:790–799
Reich BJ, Hodges JS, Zadnik V (2006) Effects of residual smoothing on the posterior of the fixed effects in disease-mapping models. Biometrics 62(4):11971206
Royle AJ, Berliner ML (1999) A hierarchical approach to multivariate spatial modeling and prediction. J Agric Biol Environ Stat 4(1):29–56
Rue H, Held L (2005) Gaussian Markov random fields: theory and applications. Chapman & Hall, New York
Sain SR, Cressie N (2007) A spatial analysis of multivariate lattice data. J Econom 140:226–259
Sain SR, Furrer R, Cressie N (2011) A spatial analysis of multivariate output from regional climate models. Ann Appl Stat 5(1):150–175
Schmidt AM, Gelfand AE (2003) A Bayesian coregionalization approach for multivariate pollutant data. J Geophys Res 108 STS 10:1–9
Spiegelhalter DJ, Best NG, Carlin BP, van der Linde A (2002) Bayesian measures of model complexity and fit (with discussion). J R Stat Soc Ser B 64:583–640
Spiegelhalter DJ, Thomas A, Best NG, Lunn D (2007) WinBUGS User Manual, Version 1.4.3
Sun D, Tsutakawa RK, Speckman PL (1999) Posterior distribution of hierarchical models using CAR(1) distributions. Biometrika 86:341–350
Sun W, Le ND, Zidek JV, Burnett R (1998) Assessment of a Bayesian multivariate interpolation approach for health impact studies. Environmetrics 9:565–586
Ugarte MD, Adin A, Goicoa T (2017) One-dimensional, two-dimensional, and three dimensional B-splines to specify space–time interactions in Bayesian disease mapping: model fitting and model identifiability. Spat Stat 22(2):451–468
Wackernagel H (2003) Multivariate geostatistics: an introduction with applications. Springer, Berlin
Wakefield J, Salway R (2001) A statistical framework for ecological and aggregate studies. J R Stat Soc Ser A 164(1):119–137
Watanabe S (2010) Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular leaning theory. J Mach Learn Res 11:3571–3594
Zhang H (2007) Maximum-likelihood estimation for multivariate spatial linear coregionalization models. Environmetrics 18:125–139
Zhang L, Baladandayuthapani V, Zhu H, Baggerly KA, Majewski T, Czerniak BA, Morris JS (2016) Functional CAR models for large spatially correlated functional datasets. J Am Stat Assoc 111(514):772–786
Acknowledgements
I am grateful for the comments of two referees, an associate editor, and the editor. Their input led me to a deeper study into the field and a major rewrite of an early version of the manuscript. This research is supported by a grant from the Natural Sciences and Engineering Research Council of Canada.
Author information
Authors and Affiliations
Corresponding author
Additional information
This invited paper is discussed in comments available at https://doi.org/10.1007/s11749-018-0606-2; https://doi.org/10.1007/s11749-018-0607-1; https://doi.org/10.1007/s11749-018-0609-z.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
MacNab, Y.C. Some recent work on multivariate Gaussian Markov random fields. TEST 27, 497–541 (2018). https://doi.org/10.1007/s11749-018-0605-3
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11749-018-0605-3
Keywords
- Asymmetric cross-spatial dependencies
- Conditional autoregressive model
- Decomposition
- Disease mapping
- Entanglement
- Gaussian Markov random fields
- Linear model of coregionalization
- Locally adaptive
- Positivity
- Separation
- Spatially varying coregionalization